The rate of magnetization of a ferromagnetic material abruptly increases at a temperature in the vicinity of its Curie point (T.degree.C.), and therefore, it is easily magnetized even in a weak signal magnetic field. By cooling it directly, the magnetization can be fixed. Thermoremanent duplication is a technique developed by applying this theory, and has already been known in the art. Usually, the thermoremanent duplication process is carried out by heating a slave tape to a temperature above the Curie point of the ferromagnetic material contained in its magnetic layer, removing the heat source, then bringing the slave tape into intimate contact with a master tape, and cooling the slave tape to a temperature below the Curie point while maintaining this intimate contact. When the slave tape is then separated from the master tape, a magnetic signal of the same shape as in the master tape is recorded in the magnetic matrix of the slave tape.
Various materials can be used as a base film of the slave tape, but actually, polyester films are used. When a magnetic recording tape based on a polyester film is used as a slave tape, and subjected to the above-mentioned thermoremanent duplication, the slave tape is placed under some tension when heated to a temperature above the Curie point of the magnetic material and then brought into intimate contact with the master tape. Owing to this stretching tension under heating, the polyester base film stretches, although to a slight degree, and as a result, strain occurs in the slave tape. The extent of the strain depends mainly upon the tension to be exerted on the base film at the time of thermoremanent duplication. Generally, slave tapes which have thus developed strain do not return to the original length within a short period of time, but some strain still remains there even when the slave tape is allowed to cool to room temperature after the thermoremanent duplication step. This residual strain decreases very slowly at room temperature, and therefore the length of the thermoremanently duplicated slave tape becomes shorter with time at a very slow rate. Such a slow change in dimension of the slave tape, even to a slight extent presents serious problems especially when it is used for video magnetic recording. When video slave tapes are commercially produced by the thermoremanent duplication method using a master tape, a playing test of the slave tape must be carried out after a lapse of a certain period of time from the thermoremanent duplication in order to examine whether normal images are recorded in the slave tape. Even if the reproduction of normal images is confirmed at this time any subsequent change in the dimension of the slave tape ascribable to the reduction of the residual strain of the slave tape would cause skew in the reproduced images corresponding to the extent of the dimensional change of the tape when the user reproduces the tape after a long period of time. Such a skew greatly impairs the commercial value of the slave tape. Dimensional changes of the slave tape of course occur even in a time interval between the thermoremanent duplication and the playing test, but can be compensated by suitably controlling the thermoremanent duplication step. Accordingly normal images can be recorded at the time of the playing test, and therefore, this does not pose a serious problem. What matters is a dimensional change in slave tapes which occurs after the playing test. This troublesome dimensional change could be avoided by providing a tape-tension controller or skew controller in a video tape recorder, but this adds to the cost of the reproducing apparatus. It is very desirable therefore to maintain normal images ensured by the playing test without resorting to such means. In order to realize such a desire, it is necessary to minimize the dimensional change that may occur after the playing test so that no substantial skew occurs in the reproduced images at whatever time the user reproduces the slave tape. With the conventional thermoremanent duplication processes, there has been no way to avoid the dimensional change of the slave tape which proceeds slowly over very long periods of time after the thermoremanent duplication. However, if the playing test for quality examination is carried out after a very long period of time, for example, one week or 10 days, from the end of the thermoremanent duplication, the dimensional change of the slave tape that occurs after the playing test would be able to be reduced to an extent sufficient to cause no detrimental skew to the reproduced images. However the lengthening of the time duration between the thermoremanent duplication step and the playing test in this manner is not only uneconomical, but also makes it difficult to control the step of producing slave tapes.